AioCare’s R&D team has created an AI algorithm for automatic detection of cough during spirometry testing. It was possible thanks to using neural networks and machine learning technology. The article has been published in the Informatics in Medicine Unlocked. To build the algorithm a considerable data set (about 20,000 spirometry records) was used, obtaining 93% detection accuracy. It is the first open-source article on how to detect coughing from an airflow signal so thoroughly described in the literature.

Nowadays, artificial intelligence becomes more and more involved in clinical diagnosis and treatment. Of course, It is also the case in lung function testing. A recent publication outlines the details on how a mobile spirometry system can detect cough during the spirometry. This efficient automatic cough detection will be of use in home monitoring and in clinical practice assisting the physician in making medical decisions. AI-based solutions become essential in any unusual circumstances as the medical stuff shortage or overwork.

In this article, the authors give a short introduction to cough detection efforts that were undertaken during the last decade and describe the solution for automatic cough detection developed for the AioCare portable spirometry system. In contrast to more popular analysis of sound and audio recordings, the solution is entirely based on airflow signals only. The analytical methodology developed in this study detects cough events during spirometry measurement automatically and with high accuracy. The solution presented in this publication is the first fully reproducible description of the automatic cough detection algorithm based totally on airflow signals. 

Read the full publication here https://www.sciencedirect.com/science/article/pii/S2352914820300587